Hi Amiga community
I just solved an issue with a Picasso 2 RTG card (Version 1.4) and quite embarassed myself, so I thougt, I'd write down what happened - so at least it could help someone else in the future having the same issue. I didn't find info about this online, so it could be that it isn't THAT obvious, but please DO let me know, if this issue is common knowledge, so I can be ashamed even more about my newbieness...
So what happened? I bought a Picasso 2 RTG Card on AmiBay recently and tested it in my A2000 as soon as it arrived. The P96 software installed normal and detected the card as expected. When I first switched to RTG mode, I heard the iconic CLICK of the relais but there was no picture. Or was it? A REALLY faint and extremely dark picture was visible, when I turned down the ambient light - so the card seemed to work, but the picture was so dark, it could hardly be seen.
I inspected the card and it looked fine, but I realized, that the RGB lines were filtered through electrolytic caps which lie in series with the trace. They did not leak, but they could be dried out, I thought - so I recapped the board. Did not help: Same issue after recap.
Now it was time to download the datasheet of the VGA chip to research, how the brightness is generated on the analog RGB lines. I soon found out, that there was a reference Input (IREF), that was supposed to have a reference current (current - NOT voltage) and this ref. current was generated by a LM317LZ voltage regulator. So maybe this IC has gone bad, was my next thought. I measured the current on the IREF pin, which is supposed to be around 6.7 mAmps - it was below 1 mA what explained the dark video output. So either the IC or the passives which set the right current must have gone bad, I thought - let's replace them. Fortunately, in hindsight, I did not have a replacement IC in my inventory, so I thought, I would look further into the circuit before ordering on Mouser, so I could order the right caps and resistors with the IC.
The circuit looked much like the reference implementation in the "GD542x Technical Reference Manual", but it had different resistor values and had some diffs that made me curious, so I began to trace the connections to the LM317. It looks something like this:
The trace to the input pin crosses the whole board, coming from pin 8 of the Zorro 2 connector, which is the minus 5V rail. Some of you will probably understand already, what happened - but I was still clueless. I just thought "minus 5 Volt? That's strange, why would they use minus 5V?" To make sure, that it was indeed -5V, I decided to measure the voltage on pin 2 of the LM317 while the Amiga was running - and measured something around +4V. So what could be wrong? The circuit between pins 2 and 3 of the LM317 seems to be closed, I did not see any other voltage input than the -5V rail - so where did the 4V come from?
I pulled out the card, switched on the Amiga and measured the -5V rail again, to make sure, that it got the right voltage from the PSU. It did NOT of course, because I run this A2000 with a replacement ATX PSU that lacks the -5V rail
So this is the easy and in hindsight obvious solution: The Picasso 2 is one of the few Zorro-Cards, that actually DOES use the -5V rail and will not work without. I haven't installed a converter with a -5V generator in this Amiga yet, because I've never ever missed the -5V before and I've completely forgotten about it... Since installing an ATX to AT converter with -5V generator between my PSU and the Amiga, the Picasso 2 works flawlessly, finally. All is well if it ends well, I think...
As using an ATX PSU gets more and more common, as old amiga PSUs fail, other people could run into this issue - even more so, as only very few cards will miss the -5V rail. So many of us can go quite far with a passive adapter only and forget about the -5V. Also, I wouldn't have expected an RTG card to use the -5V - that's why I thought, I'd share this with the community.
I just solved an issue with a Picasso 2 RTG card (Version 1.4) and quite embarassed myself, so I thougt, I'd write down what happened - so at least it could help someone else in the future having the same issue. I didn't find info about this online, so it could be that it isn't THAT obvious, but please DO let me know, if this issue is common knowledge, so I can be ashamed even more about my newbieness...
So what happened? I bought a Picasso 2 RTG Card on AmiBay recently and tested it in my A2000 as soon as it arrived. The P96 software installed normal and detected the card as expected. When I first switched to RTG mode, I heard the iconic CLICK of the relais but there was no picture. Or was it? A REALLY faint and extremely dark picture was visible, when I turned down the ambient light - so the card seemed to work, but the picture was so dark, it could hardly be seen.
I inspected the card and it looked fine, but I realized, that the RGB lines were filtered through electrolytic caps which lie in series with the trace. They did not leak, but they could be dried out, I thought - so I recapped the board. Did not help: Same issue after recap.
Now it was time to download the datasheet of the VGA chip to research, how the brightness is generated on the analog RGB lines. I soon found out, that there was a reference Input (IREF), that was supposed to have a reference current (current - NOT voltage) and this ref. current was generated by a LM317LZ voltage regulator. So maybe this IC has gone bad, was my next thought. I measured the current on the IREF pin, which is supposed to be around 6.7 mAmps - it was below 1 mA what explained the dark video output. So either the IC or the passives which set the right current must have gone bad, I thought - let's replace them. Fortunately, in hindsight, I did not have a replacement IC in my inventory, so I thought, I would look further into the circuit before ordering on Mouser, so I could order the right caps and resistors with the IC.
The circuit looked much like the reference implementation in the "GD542x Technical Reference Manual", but it had different resistor values and had some diffs that made me curious, so I began to trace the connections to the LM317. It looks something like this:
The trace to the input pin crosses the whole board, coming from pin 8 of the Zorro 2 connector, which is the minus 5V rail. Some of you will probably understand already, what happened - but I was still clueless. I just thought "minus 5 Volt? That's strange, why would they use minus 5V?" To make sure, that it was indeed -5V, I decided to measure the voltage on pin 2 of the LM317 while the Amiga was running - and measured something around +4V. So what could be wrong? The circuit between pins 2 and 3 of the LM317 seems to be closed, I did not see any other voltage input than the -5V rail - so where did the 4V come from?
I pulled out the card, switched on the Amiga and measured the -5V rail again, to make sure, that it got the right voltage from the PSU. It did NOT of course, because I run this A2000 with a replacement ATX PSU that lacks the -5V rail
So this is the easy and in hindsight obvious solution: The Picasso 2 is one of the few Zorro-Cards, that actually DOES use the -5V rail and will not work without. I haven't installed a converter with a -5V generator in this Amiga yet, because I've never ever missed the -5V before and I've completely forgotten about it... Since installing an ATX to AT converter with -5V generator between my PSU and the Amiga, the Picasso 2 works flawlessly, finally. All is well if it ends well, I think...
As using an ATX PSU gets more and more common, as old amiga PSUs fail, other people could run into this issue - even more so, as only very few cards will miss the -5V rail. So many of us can go quite far with a passive adapter only and forget about the -5V. Also, I wouldn't have expected an RTG card to use the -5V - that's why I thought, I'd share this with the community.
